ENSCORESW-2818: remove vega from xref code

[ENSCORESW-2820] Update links in pull request template

registry_public_cores.t now uses frozen test database

...the difference being that now we use the database server configured
in MultiTestDB.conf rather than the public one. Note that since the
Registry scans all databases (modulo version number) on the server and
does not know about test prefixes, there might be some warnings
about duplicate species/group combinations; these should ultimately be
harmless though and as long as the test-database schema version matches
that of the API, the tests should succeed.

Had to rewrite the "registry knows more than 10 DBAs" test to "registry
knows at least 1 DBA" because by default that is what MultiTestDB
provides.

See-also: https://www.ebi.ac.uk/panda/jira/browse/ENSCORESW-2325

backport ENSCORESW-2767: multi-species support for protein feature

ENSCORESW-2810: pass RFAMParser even if no xrefs found

Do not call fetchrow_hashref() if the statement has been exhausted

update to latest schema

ENSCORESW-2740: fetch correct xref

The Compara API can deal with Slice objects directly

This also gets rid of the deprecation warning of
fetch_all_by_GenomeDB_region

fixed cpanfile syntax

ENSCORESW-2758: default align type is NULL

Bugfix/reactome acc

SequenceOntologyMapper deprecation

Pre release 94 cleanup

ENSCORESW-2695: expand methods for REST support

fix patch name

Apply missing MySQL test-database patches from 90 to 93